athena 0.0.1.53
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/COPYING +676 -0
- data/ChangeLog +5 -0
- data/README +33 -0
- data/Rakefile +23 -0
- data/bin/athena +165 -0
- data/example/config.yaml +72 -0
- data/example/example.xml +26 -0
- data/example/sisis-ex.txt +90 -0
- data/lib/athena.rb +68 -0
- data/lib/athena/formats.rb +77 -0
- data/lib/athena/formats/dbm.rb +65 -0
- data/lib/athena/formats/sisis.rb +83 -0
- data/lib/athena/formats/xml.rb +278 -0
- data/lib/athena/parser.rb +90 -0
- data/lib/athena/record.rb +97 -0
- data/lib/athena/util.rb +50 -0
- data/lib/athena/version.rb +55 -0
- metadata +81 -0
@@ -0,0 +1,77 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
module Athena
|
30
|
+
|
31
|
+
class Formats
|
32
|
+
|
33
|
+
@formats = { :in => {}, :out => {} }
|
34
|
+
|
35
|
+
class << self
|
36
|
+
|
37
|
+
def formats
|
38
|
+
Formats.instance_variable_get :@formats
|
39
|
+
end
|
40
|
+
|
41
|
+
def [](direction, format)
|
42
|
+
formats[direction][format]
|
43
|
+
end
|
44
|
+
|
45
|
+
def valid_format?(direction, format)
|
46
|
+
formats[direction].has_key? format
|
47
|
+
end
|
48
|
+
|
49
|
+
def convert(*args)
|
50
|
+
raise NotImplementedError, 'must be defined by sub-class'
|
51
|
+
end
|
52
|
+
|
53
|
+
private
|
54
|
+
|
55
|
+
def register_format(direction, format)
|
56
|
+
formats[direction][format] = self
|
57
|
+
end
|
58
|
+
|
59
|
+
def register_formats(direction, *formats)
|
60
|
+
formats.each { |format|
|
61
|
+
register_format(direction, format)
|
62
|
+
}
|
63
|
+
end
|
64
|
+
|
65
|
+
end
|
66
|
+
|
67
|
+
def parse(*args)
|
68
|
+
raise NotImplementedError, 'must be defined by sub-class'
|
69
|
+
end
|
70
|
+
|
71
|
+
end
|
72
|
+
|
73
|
+
end
|
74
|
+
|
75
|
+
Dir[__FILE__.sub(/\.rb$/, '/**/*.rb')].each { |rb|
|
76
|
+
require rb
|
77
|
+
}
|
@@ -0,0 +1,65 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
require 'iconv'
|
30
|
+
|
31
|
+
module Athena
|
32
|
+
|
33
|
+
class Formats
|
34
|
+
|
35
|
+
class DBM < Athena::Formats
|
36
|
+
|
37
|
+
register_formats :out, 'dbm', 'midos'
|
38
|
+
|
39
|
+
CRLF = "\015\012"
|
40
|
+
|
41
|
+
ICONV_TO_LATIN1 = Iconv.new('latin1', 'utf-8')
|
42
|
+
|
43
|
+
def self.convert(record)
|
44
|
+
dbm = ["ID:#{record.id}"]
|
45
|
+
record.struct.each { |field, struct|
|
46
|
+
strings = struct[:elements].inject([]) { |array, element|
|
47
|
+
values = (struct[:values][element] || []).map { |v|
|
48
|
+
(v || '').strip.gsub(/(?:\r?\n)+/, ' ')
|
49
|
+
}.reject { |v| v.empty? }
|
50
|
+
|
51
|
+
array << (values.empty? ? struct[:empty] : values.join('|'))
|
52
|
+
}
|
53
|
+
|
54
|
+
dbm << "#{field.to_s.upcase}:#{ICONV_TO_LATIN1.iconv(struct[:string] % strings)}"
|
55
|
+
}
|
56
|
+
dbm << '&&&'
|
57
|
+
|
58
|
+
dbm.join(CRLF) << CRLF << CRLF
|
59
|
+
end
|
60
|
+
|
61
|
+
end
|
62
|
+
|
63
|
+
end
|
64
|
+
|
65
|
+
end
|
@@ -0,0 +1,83 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
module Athena
|
30
|
+
|
31
|
+
class Formats
|
32
|
+
|
33
|
+
class Sisis < Athena::Formats
|
34
|
+
|
35
|
+
register_format :in, 'sisis'
|
36
|
+
|
37
|
+
attr_reader :record_element, :config, :parser
|
38
|
+
|
39
|
+
def initialize(parser)
|
40
|
+
config = parser.config.dup
|
41
|
+
|
42
|
+
case @record_element = config.delete(:__record_element)
|
43
|
+
when String
|
44
|
+
# fine!
|
45
|
+
when nil
|
46
|
+
raise NoRecordElementError, 'no record element specified'
|
47
|
+
else
|
48
|
+
raise IllegalRecordElementError, "illegal record element #{@record_element}"
|
49
|
+
end
|
50
|
+
|
51
|
+
@config = config
|
52
|
+
@parser = parser
|
53
|
+
end
|
54
|
+
|
55
|
+
def parse(source)
|
56
|
+
record = nil
|
57
|
+
|
58
|
+
source.each { |line|
|
59
|
+
element, value = line.match(/(\d+).*?:\s*(.*)/)[1, 2]
|
60
|
+
|
61
|
+
case element
|
62
|
+
when record_element
|
63
|
+
record.close if record
|
64
|
+
record = Athena::Record.new(parser.block, value)
|
65
|
+
else
|
66
|
+
record.update(element, value, config[element])
|
67
|
+
end
|
68
|
+
}
|
69
|
+
|
70
|
+
record.close if record
|
71
|
+
end
|
72
|
+
|
73
|
+
class NoRecordElementError < StandardError
|
74
|
+
end
|
75
|
+
|
76
|
+
class IllegalRecordElementError < StandardError
|
77
|
+
end
|
78
|
+
|
79
|
+
end
|
80
|
+
|
81
|
+
end
|
82
|
+
|
83
|
+
end
|
@@ -0,0 +1,278 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
require 'forwardable'
|
30
|
+
|
31
|
+
require 'rubygems'
|
32
|
+
|
33
|
+
require 'xmlstreamin'
|
34
|
+
require 'nuggets/hash/insert'
|
35
|
+
|
36
|
+
module Athena
|
37
|
+
|
38
|
+
class Formats
|
39
|
+
|
40
|
+
class XML < Athena::Formats
|
41
|
+
|
42
|
+
include Util
|
43
|
+
|
44
|
+
register_format :in, 'xml'
|
45
|
+
|
46
|
+
attr_reader :spec, :listener
|
47
|
+
|
48
|
+
def initialize(parser)
|
49
|
+
@spec = build_spec(parser)
|
50
|
+
@listener = XMLStreamin::XMLStreamListener.new(@spec)
|
51
|
+
end
|
52
|
+
|
53
|
+
def parse(source)
|
54
|
+
REXML::Document.parse_stream(source, listener)
|
55
|
+
end
|
56
|
+
|
57
|
+
private
|
58
|
+
|
59
|
+
def build_spec(parser)
|
60
|
+
config = parser.config.dup
|
61
|
+
|
62
|
+
case record_element = config.delete(:__record_element)
|
63
|
+
when String
|
64
|
+
# fine!
|
65
|
+
when nil
|
66
|
+
raise NoRecordElementError, 'no record element specified'
|
67
|
+
else
|
68
|
+
raise IllegalRecordElementError, "illegal record element #{record_element}"
|
69
|
+
end
|
70
|
+
|
71
|
+
element_specs = config.inject({}) { |specs, (element, element_spec)|
|
72
|
+
element_spec.each { |field, c|
|
73
|
+
element.split('/').reverse.inject({}) { |hash, part|
|
74
|
+
s = define_spec(element, field, c, hash.empty? ? :default : hash)
|
75
|
+
merge_specs(hash, part, s)
|
76
|
+
}.each { |key, s|
|
77
|
+
merge_specs(specs, key, s)
|
78
|
+
}
|
79
|
+
}
|
80
|
+
|
81
|
+
specs
|
82
|
+
}
|
83
|
+
|
84
|
+
record_spec = RecordSpec.new(parser)
|
85
|
+
record_spec.specs!(element_specs)
|
86
|
+
|
87
|
+
root_spec = BaseSpec.new
|
88
|
+
root_spec.specs!(record_element => record_spec)
|
89
|
+
|
90
|
+
spec = BaseSpec.new
|
91
|
+
spec.default!(root_spec)
|
92
|
+
|
93
|
+
verbose(:spec, BaseSpec) do
|
94
|
+
spec.inspect_spec
|
95
|
+
end
|
96
|
+
|
97
|
+
spec
|
98
|
+
end
|
99
|
+
|
100
|
+
def define_spec(element, field, config, arg)
|
101
|
+
spec = ElementSpec.new(element, field, config)
|
102
|
+
|
103
|
+
case arg
|
104
|
+
when Hash
|
105
|
+
spec.specs!(arg)
|
106
|
+
else
|
107
|
+
spec.default!(SubElementSpec.new(spec))
|
108
|
+
end
|
109
|
+
|
110
|
+
spec
|
111
|
+
end
|
112
|
+
|
113
|
+
def merge_specs(container, key, spec)
|
114
|
+
container.insert!(key, spec) { |s1, s2|
|
115
|
+
if s1.respond_to?(:specs!)
|
116
|
+
s1.specs!(s2.respond_to?(:specs) ? s2.specs : s2)
|
117
|
+
s1
|
118
|
+
else
|
119
|
+
s1.merge(s2)
|
120
|
+
end
|
121
|
+
}
|
122
|
+
end
|
123
|
+
|
124
|
+
class BaseSpec < XMLStreamin::XMLSpec
|
125
|
+
|
126
|
+
include Util
|
127
|
+
|
128
|
+
@level = 0
|
129
|
+
|
130
|
+
def start(context, name, attrs)
|
131
|
+
verbose(:xml) do
|
132
|
+
spit "#{indent(level)}<#{name}>"
|
133
|
+
step :down
|
134
|
+
|
135
|
+
attrs.each { |attr|
|
136
|
+
spit "#{indent(level + 1)}[#{attr[0]} = #{attr[1]}]"
|
137
|
+
}
|
138
|
+
end
|
139
|
+
|
140
|
+
return context
|
141
|
+
end
|
142
|
+
|
143
|
+
def text(context, data)
|
144
|
+
verbose(:xml) do
|
145
|
+
content = data.strip
|
146
|
+
spit "#{indent(level)}#{content}" unless content.empty?
|
147
|
+
end
|
148
|
+
|
149
|
+
return context
|
150
|
+
end
|
151
|
+
|
152
|
+
def done(context, name)
|
153
|
+
verbose(:xml) do
|
154
|
+
step :up
|
155
|
+
spit "#{indent(level)}</#{name}>"
|
156
|
+
end
|
157
|
+
|
158
|
+
return context
|
159
|
+
end
|
160
|
+
|
161
|
+
def empty(context)
|
162
|
+
verbose(:xml) do
|
163
|
+
step :up
|
164
|
+
end
|
165
|
+
|
166
|
+
return context
|
167
|
+
end
|
168
|
+
|
169
|
+
def inspect_spec(element = nil, level = 0)
|
170
|
+
if respond_to?(:field)
|
171
|
+
msg = "#{indent(level)}[#{element}] #{field.to_s.upcase} -> #{name}"
|
172
|
+
respond_to?(:spit) ? spit(msg) : warn(msg)
|
173
|
+
specs.each { |e, s|
|
174
|
+
s.inspect_spec(e, level + 1)
|
175
|
+
}
|
176
|
+
else
|
177
|
+
if specs.empty?
|
178
|
+
specs.default.inspect_spec('?', level)
|
179
|
+
else
|
180
|
+
specs.each { |e, s|
|
181
|
+
s.inspect_spec(e, level)
|
182
|
+
}
|
183
|
+
end
|
184
|
+
end
|
185
|
+
end
|
186
|
+
|
187
|
+
private
|
188
|
+
|
189
|
+
def level
|
190
|
+
BaseSpec.instance_variable_get :@level
|
191
|
+
end
|
192
|
+
|
193
|
+
def step(direction)
|
194
|
+
steps = { :down => 1, :up => -1 }
|
195
|
+
BaseSpec.instance_variable_set :@level, level + steps[direction]
|
196
|
+
end
|
197
|
+
|
198
|
+
end
|
199
|
+
|
200
|
+
class RecordSpec < BaseSpec
|
201
|
+
|
202
|
+
attr_reader :parser
|
203
|
+
attr_accessor :record
|
204
|
+
|
205
|
+
def initialize(parser)
|
206
|
+
super()
|
207
|
+
|
208
|
+
@parser = parser
|
209
|
+
end
|
210
|
+
|
211
|
+
def start(context, name, attrs)
|
212
|
+
super
|
213
|
+
|
214
|
+
self.record = Athena::Record.new(parser.block)
|
215
|
+
end
|
216
|
+
|
217
|
+
def done(context, name)
|
218
|
+
super
|
219
|
+
|
220
|
+
record.close
|
221
|
+
end
|
222
|
+
|
223
|
+
end
|
224
|
+
|
225
|
+
class ElementSpec < BaseSpec
|
226
|
+
|
227
|
+
attr_reader :name, :field, :config
|
228
|
+
attr_accessor :record
|
229
|
+
|
230
|
+
def initialize(name, field, config)
|
231
|
+
super()
|
232
|
+
|
233
|
+
@name = name
|
234
|
+
@field = field
|
235
|
+
@config = config
|
236
|
+
end
|
237
|
+
|
238
|
+
def start(context, name, attrs)
|
239
|
+
super
|
240
|
+
|
241
|
+
self.record = Athena::Record[field, config]
|
242
|
+
end
|
243
|
+
|
244
|
+
def text(context, data)
|
245
|
+
super
|
246
|
+
|
247
|
+
record.update name, data
|
248
|
+
end
|
249
|
+
|
250
|
+
end
|
251
|
+
|
252
|
+
class SubElementSpec < BaseSpec
|
253
|
+
|
254
|
+
extend Forwardable
|
255
|
+
|
256
|
+
# Forward to parent element; need to specify *all* its attributes and methods
|
257
|
+
def_delegators :@parent, :name, :field, :config, :record, :start, :text
|
258
|
+
|
259
|
+
def initialize(parent)
|
260
|
+
super()
|
261
|
+
|
262
|
+
@parent = parent
|
263
|
+
default! self
|
264
|
+
end
|
265
|
+
|
266
|
+
end
|
267
|
+
|
268
|
+
class NoRecordElementError < StandardError
|
269
|
+
end
|
270
|
+
|
271
|
+
class IllegalRecordElementError < StandardError
|
272
|
+
end
|
273
|
+
|
274
|
+
end
|
275
|
+
|
276
|
+
end
|
277
|
+
|
278
|
+
end
|